Objective.Function: Objective Function
In raff-k/Lslide: Functions for object-oriented image analysis in landslide science

Objective.Function

R Documentation

Objective Function

Description

The Objective Function by ESPINDOLA et al. (2006) finds optimal scale parameters by the comparision of the sum of the normalized local variance and the normalized Moran's I of different segmented images.

Usage

Objective.Function(Tool, Scale.Input.Grid,
  Scale.Input.Grid.Cell.Size = NULL, Scale.Statistic.Min.Size = "0",
  Objective.Function.save = FALSE, Objective.Function.save.path = NULL,
  Count = "1", Min = "0", Max = "0", Range = "0", Sum = "0",
  Mean = "1", Var = "1", Stddev = "0",
  Objective.Function.save.name = "",
  Objective.Function.Mean.Segmentation.Grid = paste0(tmp.path,
  "Objective.Function.Mean.Segmentation.Grid.sgrd"),
  Objective.Function.Count.Grid = paste0(tmp.path,
  "Objective.Function.Count.Grid.sgrd"),
  Objective.Function.MoransI = paste0(tmp.path,
  "Objective.Function.MoransI.txt"), Quantile = 0, Scales,
  Grass.Objective.Function.Method = "Threshold", Segments.Poly,
  Segments.Grid, Grass.Segmentation.Minsize = NA,
  Grass.Segmentation.Threshold = NA, Seed.Method = "",
  env = RSAGA::rsaga.env(), show.output.on.console = FALSE,
  quiet = TRUE, HiPassFilter.input.segmentation = NULL,
  HiPassFilter.input.filter = Scale.Input.Grid,
  HiPassFilter.scale.factor = NULL, HiPassFilter.threshold = NULL,
  do.storeGrids = FALSE, ...)

Arguments

`Tool`	open-source software to compute segmentation analysis. GRASS or SAGA
`Scale.Input.Grid`	input grid for computing segmentation scale parameters
`Scale.Input.Grid.Cell.Size`	cell size of input grid. Default: prod(raster::res(raster::raster(slp.tif.path)))
`Scale.Statistic.Min.Size`	min size of area/polygon which is included in statistics (usefull for SAGA GIS segmentations). Default: "0"
`Objective.Function.save`	save estimations of function. Default: FALSE
`Objective.Function.save.path`	path where estimations are stored. Default: input path of segment.poly
`Count`	amount of cells (Grid Statistics). Default:"1"
`Min`	minimum value (Grid Statistics). Default:"0"
`Max`	maximum value (Grid Statistics). Default:"0"
`Range`	range of values (Grid Statistics). Default:"0"
`Sum`	sum of values (Grid Statistics). Default:"0"
`Mean`	mean of values (Grid Statistics). Default:"1"
`Var`	variance of values (Grid Statistics). Default:"1"
`Stddev`	standard deviation (Grid Statistics). Default:"0"
`Objective.Function.save.name`	name for file. Default: ""
`Objective.Function.Mean.Segmentation.Grid`	- output grid with mean values for segments. Default: temp
`Objective.Function.Count.Grid`	output grid with amount of cells in segments. Default: temp
`Objective.Function.MoransI`	output text file with Morans'I. Default: temp
`Quantile`	qunatile (Grid Statistics). Default:"0"
`Scales`	containing scale parameters for loop-segmentation
`Grass.Objective.Function.Method`	determining on what parameter the objective function (~loop) is performed. Default: "Threshold"
`Segments.Poly`	...
`Grass.Segmentation.Minsize`	= NA,
`Grass.Segmentation.Threshold`	= NA
`Seed.Method`	""
`env`	...
`HiPassFilter.input.segmentation`	path to load
`HiPassFilter.input.filter`	path to load
`HiPassFilter.scale.factor`	scale factor of high-pass filter
`HiPassFilter.scale.threshold`	threshold of high-pass filter
`other...`	see `segmentation`

Note

ESPINDOLA, G.M., CAMARA, G., REIS, I.A., BINS, L.S. & A.M. MONTEIRO (2006): Parameter selection for region-growing image segmentation algorithms using spatial autocorrelation. - International Journal of Remote Sensing 27, 14, 3035-3040.

raff-k/Lslide documentation built on March 29, 2022, 6:52 p.m.